US3431356A - Apparatus and method for reconstructing speech - Google Patents

Abstract

1,151,536. Diver's communication system. M. COPEL. 3 June, 1966 [4 June, 1965], No. 24903/66. Heading H4R. Speech signals from a diver operating in an ambient atmosphere of helium and/or high pressure are rendered more intelligible by selecting two or more separate frequency bands from the speech signal and moving them down in frequency so that they are closer together, but not overlapping. In the described embodiment, the bands selected are 850 to 1700 c.p.s. and 2400 to 5000 c.p.s. and they are modulated twice, first by a common frequency of 8000 c.p.s. and then the upper sidebands are modulated by frequencies of 8600 and 9000 c.p.s. respectively from which the lower sidebands of 250 to 1100 c.p.s. and 1400 to 4000 c.p.s. are selected, added together and reproduced. It is suggested that a lower frequency band may be transmitted unchanged. Circuits for a system in which speech signals from three divers may be treated by apparatus according to the invention in order to allow them to converse easily with one another or with a tender, are illustrated in Figs. 4a, 4b and 4c (not shown).

Description

March 4, 1969 Filed June 4, 1965 Sheet INVENTOR /H/a//aL aan March 4, 1969 MicbEL 3,431,356

APPARATUS AND METHOD FOR RECONSTRUCTING SPEECH Fewew V 55'0 /200 Z400 .5000

March 4, 1969 M. COPEI. 3,431,356

APPARATUS AND METHOD FOR RECONSTRUCTING SPEECH Filed June 4. 1965 Sheet 3 of s TEA/05 l "5 @ma HUK. Pfl/@V5.5 E I I l m3 l .L -I E :40 ma ,9105 I Rin s 'P2B/Vy E m7 d L INVENTOR. ad /W/w/m 0F54 c %a March 4,

M. COFEL.`

APPARATUS AND METHOD FOR RECONSTRUCTING SPEECH Mariah 4, 1969 M. coPEL 3,431,356 I APPARATUS AND METHOD FOR RECONSTRUCTING SPEECH Filed June 4, 1965 Sheet 5 of 5 United States Patent O 3,431,356 APPARATUS AND METHOD FOR RECONSTRUCTING SPEECH Michel Copel, Huntington, N.Y., assignor to Integrated Electronics Corporation Filed June 4, 1965, Ser. No. 461,559 U.S. Cl. 179-1 10 Claims Int. Cl. H04m 1/00; H04b J 66 ABSTRACT F THE DISCLOSURE To correct the voice distortions created by speaking using a lighter-than-air gaseous atmosphere, the distorted audio signal is converted into an electrical signal which is ltered into two specific separated frequency bands, Ithe ltered bands are shifted downwardly by different amounts and added to reconvert them to an audio output.

The invention described herein may be manufactured and used by or for the Government of the United States of America for governmental purposes without the payment of any royalties thereon or therefor.

This invention relates to the reconstruction of speech which by reason of environmental conditions on the speaker has become non-linearly distorted in frequency spectrum to such an extent as to be substantially unintelligible. More particularly the invention relates to a method and apparatus for reconstructing such substantially unintelligible speech to obtain intelligible speech.

The invention is particularly useful in the transmission of speech from divers working at a considerable depth lbelow the surface of the water and in an ambient condition including the breathing of air including a helium gas mixture. The helium gas mixture is provided to reduce the accumulation of nitrogen in the divers blood stream and thus avoid or reduce the narcosis effect otherwise occuring. Other exotic gases, such as hydrogen, can also be used for the same function.

In normal audible speech, each word has a frequency spectrum which is familiar to and understandable by humans. However, in speaking under the conditions of a helium mixture atmosphere and an abnormal pressure, such for example as the pressure at 200 feet below the surface of water, each of the above abnormal conditions operate to produce an effect on the human speech producing mechanism such that the speech frequency spectrum becomes distorted in a non-linear manner and is either unintelligible or substantially unintelligible.

This problem of inability of a diver operating under such conditions to communicate to another diver or to persons on a tender on the surface of the water and the inability of a diver to communicate from a submerged chamber under the same conditions to persons on a tender on 4the surface has seriously impeded underwater research for several years and many attempts have been made without success to overcome the problem.

In the prior art, various means and methods have been employed successfully to manipulate normal or substantially normal speech such that the processed speech, after transmission, is as close as possible to the original speech. To the best of applicants knowledge, no one has disclosed a method or apparatus for reconstructing speech such that non-linearly distorted speech spectrum can be recon- ICC structed to a speech departing from the original speech in a direction to provide an intelligible speech spectrum.

Itis thus an object of this invention to provide a method and apparatus for reconstructing speech having a substantially unintelligible speech spectrum due to non-linear speech frequency distortion into speech having an intelligible speech spectrum.

A further object is to provide a method and apparatus for reconstructing speech distorted by abnormal ambient conditions of pressure and/or exotic gas atmosphere on the speaker to provide speech of substantially improved intelligibility.

A more specific object is to provide a method and apparatus for converting speech distorted by conditions of abnormal pressure and the effects of helium gas on the speaker to provide speech of improved intelligibility.

Briefly stated applicants invention comprises the method of and means for selecting spaced frequency bands of the original speech spectrum, and shifting the bands downwardly in frequency by different degrees dependent upon the above described ambient conditions to provide speech of improved intelligibility.

Other objects and advantages will appear from the following description of an example of the invention and from the accompanying drawings in which:

FIG. l shows, in block diagram form, the essential elements of a system embodying the method and apparatus of the invention,

FIG. 2 shows, in block diagram form, the essential elements of a modied system embodying the method and apparatus of the invention,

FIG. 3 is a diagram indicating two selected band filters and the final positions of the bands reorien-ted in accordance with principles of the invention, and

FIGS. 4a, 4b and 4c are a detailed circuit of one apparatus in accordance with the invention.

Referring to FIG. l of the drawings, there is shown therein a block diagram of one electrical system indicated generally by numeral 10 and exemplifying the method and apparatus according to the invention.

In system 10 the distorted speech of the diver is translated to electrical current by a transducer 12 connected by conductor 14 to a preamplifier 16. The amplied signal is passed -by lines 18, 20 and 22 to bandpass, filters 24 and 26 which iilters are selected to develop spaced frequency bands of the original Speech spectrum. For example, referring to FIG. 3 bands B and C may be developed respectively by filters 24 and 26, band B covering frequency from 850 c. to 1700 c. and band C covering 2400 c. to 5000 c., the bands thus being initially separated by 700 c.

In accordance with the invention, the selected bands are then reoriented in frequency downwardly by unequal amounts to modify the original distorted speech spectrum and to move the bands into a relationship found in normal speech spectrum. To better understand the problem and the method and apparatus for solution of the same, it is to `be understood that the -divers speech under conditions of abnormal pressure, such for example, .as the pressure at 200 feet below sea level, and exotic gas ambient such, for example, as a mixture of 3% oxygen, 80% helium and 17% nitrogen, becomes unintelligible and the speech sounds somethin-g like an uninteligible Donald Duck-like voice. In accordance with this invention, it has been found that by selecting spaced bands of the original distorted speech and reorienting the selected bands downwardly and toward each other in amounts depending upon the percentage of helium used in the ambient atmosphere and the pressure on the diver, a substantial improvement in speech spectrum can be obtained and the speech becomes intelligible.

Accordingly, in system means are provided for accomplishing the the reorientation described above. In this respect any suitable means may be provided. In the example of system 1, FiG. 1, separate modulating and filter means are connected to the output of each of the filters 24 and 26.

Thus the output of filter 24 is fed by a line 28 to a modulator which also receives the output of an oscillator 32, through a line 34. The output of modulator 30 is fed by a line 36 to a high pass filter 38 to exclude undesired frequencies. The output of filter 38 is passed by a line 40 to a demodulator 42, which receives the output of a variable oscillator 44 through a line 46. The output of demodulator 42 is fed by a line 48 to a low pass filter to exclude undesired frequencies.

In the same manner B and C, i.e., the output of filter 26, is reoriented downward in frequency by the provision of a similar group of elements including a modulator 52 and associated oscillator 58, a high pass filter 62, a demodulator 66 and associated oscillator 70 and a low-pass filter 74. The elements are connected in the same manner as described `for the group of elements connected to filter 24. That is, the modulator 52 is connected by lines 54 and 56, respectively to filter 26 and oscillator 58. The output of modulator 52 is passed by line 60 to filter 62, thence by line 64 to demodulator 66 also connected to oscillator 70 by line 68 and finally by line 72 to filter 74.

Oscillators 32 and 58 are selected to provide a suitable beat for upward orientation of bands B and C. Suita-bly a pair of 8000 c. oscillators have been used. Oscillators 44 and 70 are selected in frequency such that the net effective shift as indicated in FIG. 3 is a downward shift of bands B and C and a movement of the bands closer together, the net shift, of each band, and the ratio of of shift of the bands being dependent upon the conditions of helium mixture and ambient pressure under which the diver is speaking.

FIG. 3 is provided to indicate visually the results obtained by operation of the above-described system. Thus bandpass filters 24 and 26 may be selected to provide respectively band B 4having frequency limits 850 c. and 1700 c. and band C having frequency limits of 2400 c. and 5000 c., bands B and C being spaced by 700 c. bands B and C are thus selected spaced portions of the original distorted speech spectrum.

`Bands B and C may be reoriented downwardly and towards each other to provide reoriented b and B of limits 250 c. and 1100 c. and reoriented band C of limits 1200 c. and 4000 c. by selecting oscillators 32 and S8 of f 8000 c. each and oscillators 44 and 70 of respectively 8600 c. and 9000 c. Thus, modulator 30, oscillator 32 and filter 38 reorient band B upwardly in frequency by 8000 c. to aband (not shown) of limits 8850 c. and 9700 c. and demodulator 42, oscillator 44 and filter 50 reorient the band downwardly in frequency by 8600 c. to the band B of limits 250 c. to 1100 c. shown in FIG. 3. In the same manner modulator 52, oscillator 58 and filter 62 reorient band C upwardly by 8000 c. to a band (not shown) of limits 10400 c. to 13,000 c. and demodulator 66, oscillator 70 and lter 74 reorient the band downwardly by 9000 c. to band C' of 1400 c. to 4000 c. shown in FIG. 3. Reoriented bands B and C are thus lower in frequncy and closer together (i.e. spaced by 300 c.) as compared to -the original distorted bands B and C.

As shown in FIG. 1 the outputs of filters 50 and 74 are passed respectively by lines 76 and S0 and lines 78 and 80 to a mixer 82 which in turn supplies the combined output by line 84 to suitable receiving equipment indicated by the head1-phone set 86.

It is to be understood that the degree of non-linear distortion of original speech varies in accordance with the percentage of helium in the atmosphere in which the speaker is located and to some extent upon the ambient pressure on the speaker. Thus, while improvement in speech intelligibility will result from the method and apparatus generally described above and involving reorientating the original selected bands downwardly and toward each other, still further improvements are obtained from selecting the oscillators 44 and 70 dependent upon the specific conditions of helium percentage and fpressure to obtain maximum improvement in speech intelligibility.

Oscillators 44 and 70 may eb non-adjustable oscillators selected for average ambient conditions and the system will provide improved intelligible speech under various ambient conditions. However, in accordance with a further aspect of this invention, manually adjustable oscillators are provided for the elements 44 and 70. One advantage of this arrangement is to provide flexibility inthe system to adjust the system to changes in speaker ambient conditions to adjust for maximum speech intelligibility. A second advantage is to enable the system operator to obtain maximum intelligibility of speech by listening to the corrected speech and adjusting the oscillators 44 and 70 while listening to improve speech intelligibility, for the specific ambient conditions and in accordance with the hearing of the operator.

The system above-described and shown in FIG. 1 was tested with satisfactory results under various mixtures of helium and at pressures corresponding to 200 feet and 300 feet sea depths. Under these conditions speech intelligibility was estimated to have increased from 20% to It is also feasible to increase intilligibility further by increasing the number of originally selected speech spectrum bands and reorienting the same in the manner described above. The practical limit on the number of bands selected at present appears to be a question of cost of the System elements and the need to maintain spaced relationship of adjacent bands to avoid intermodulation.

It is also possible, as shown in FIG. 3, to utilize a lower spectrum band without reorientation of the same. Thus, in FIG. 2 there is added a bandpass filter 100 which is connected by lines 102 and 104 between the lines 20 and 76 shown in FIG. 1 to directly feed the mixer 82. Filter is selected to provide a band below the spectrum of band B and spaced therefrom in frequency.

In considering the above-described method and apparatus, it is to be noted that currently deep sea divers communicate in pat, i.e. preformed, phrases which seriously restrict the extent of information which may be transmitted. As the operating depth increases and/ or the ratio of helium in the breathing mixture increases, even very limited communication becomes diflicult and often impossible. Where random information is needed, i.e. information to describe a condition found and which cannot be transmitted by pat preformed phrases, the speech is very frequently unintelligible and at best requires frequent repetition as well as a listener who is extremely well experienced in listening to this type of speech.

However, with the apparatus and method of this invention random speech on any subject may be transmitted and intelligibly received. Ihe significance of the subject device and invention is that for the first time continuous random speech communication is possible between sea divers and between the divers and the tender. Since voice communication is essential in deep sea diving operations as, for example, exploration, research, rescue, and salvage, and since the physical equipment for maintaining a diver under deep diving conditions for long periods of time has been available for some years, the subject invention now provides the essential complement of intelligible voice communication to provide safe and efficient operation of deep sea divers.

Referring to FIGS. 4a, 4b and 4c, there is shown in detail the elements and arrangement of one exemplary circuit incorporating the subject invention and which has been tested and found to provide excellent results in irnproving the intelligibility speech distorted by ambient conditions of helium mixture and abnormal pressure. Values of the elements shown in FIGS. 4a, 4b and 4c are indicated in the following table:

Component values for the elements shown in FIGS. 4a, 4b and 4c.

R1 18K R66 56K R2 10-33 R67 3.3K R3 12K R68 3.9K R4 1.2K R69 10K R5 1.50K R70 3.3K R6 100 R71 56K R7 390 R72 18K R8 56K R73 10-33 R9 3.3K R74 12K R10 3.9K R75 1.2K R11 56K R76 100 R12 750 R77 390 R13 560 R78 150K R15 1.8K R79 10 R16 2.7K R80 2.2K R17 22K R81 1K R18 470K R82 2.2K R19 62 R83 2.2K R20 2.7K R84 1K R21 1.2K R85 2.2K R22 27 R86 1K R23 12K R87 33K R24 6.8K R88 10K R25 6.8K R89 2.2K R26 1K R90 1K R27 1K R91 2.2K R28 2.2K R92 2.2K R29 1K R93 1K R30 2.2K R94 2.2K R31 2.2K R95 4.7K R32 1K R96 10K R33 2.2K R97 33K R34 1K R98 1K R35 33K R99 6.8K R36 10K R100 56K R37 4.7K R101 750 R38 2.2K R102 750 R39 1K R103 560 R40 22K R104 2.1K R41 2 2K R105 1.8K R42 1K R106 62 R43 2.2K R107 2.7K R44 6.8K R108 22K R45 1K R109 470K R46 33K R110 1.2K R47 4 71( R111 27 R48 10K R112 12K R49 150K R113 680 R50 18K R114 1K R51 10-33 R115 330 R52 12K R116 100 R53 1,21( R117 22K R54 100 R118 33 R55 390 R119 200 R56 3.3K R120 1 R57 56K R121 470 R58 3.9K R122 1K R59 18K R123 1K R60 10-33 R124 1K R61 150K R125 1K R62 12K R126 560 R63 1.2K R127 560 R64 100 R128 560 R65 390 R200 220 6 Components values for the element shown in FIGS. 4a, 4b and llc-Conzinuetl No value 4. 7

Q1 300C022-1 Q11 300C022-1 Q2 2N526 Q12 300C022-1 Q3 300C022-l Q13 300C022-1 Q4 2N526 Q14 300C022-1 Q5 300C0221 Q15 2N465 Q6 2N526 Q16 2N539A Q7 300C022-1 Q17 2N43A Q8 2N526 Q18 2N43A Q9 2N465 Q19 2N539A Q10 2N539A Q20 2N539A It will be understood that various changes in the details, materials and arrangements of parts and steps which have been herein described and illustrated in order to explain the nature of the invention, may be made by those skilled in the art within the principle and scope of the invention as expressed in the appended claims.

I claim:

1. The method of improving the intelligibility of underwater speech communication originally unintelligible because of frequency distortion due to ambient conditions, on the speaker, of lighter than air gaseous atmosphere f and greater than atmospheric pressure at a deeply submerged location comprising the steps of:

converting said unintelligible speech into an analog electrical signal, filtering said analog speech spectrum to select at least two spaced bands within the range of 85() to 6,500 cycles having extremely sharp skirts containing a substantial portion of the original unintelligible speech,

then linearly shifting said bands downwardly in frequency by different degrees of shift, the higher band being shifted by a greater amount, selected dependent upon said ambient conditions and in a relation to move said bands closer together, but still spaced apart,

adding the outputs from said shifted bands, and

,converting said added outputs to an audio signal to obtain intelligible speech.

2. The method as set forth in claim 1, wherein the width of said selected bands and relative degree of shift downwardly are selected to provide an ultimate spaced relationship of the adjusted bands such that intermodulation between the adjusted bands is avoided.

3. The method as set forth in claim 1, wherein the second mentioned step is accomplished manually such that the degree of downward shift of said bands may be controlled by ear tuning of said downward shifts to obtain speech of maximum intelligib'ility.

4. The method of claim 1 in which as said third mentioned step said bands are rst shifted upwardly in frequency and thereafter downwardly in frequency, the net shift of each band being downward in degree dependent upon said ambient conditions and in a relation of relative band shifts to move said bands closer together in frequency than the original spread between said selected bands.

5. The method as set forth in claim 4, wherein the width of said selected bands and relative degree of shift downwardly are selected to provide an ultimate spaced relationship of the adjusted bands such that intermodulation between the adjusted bands is avoided.

6. The method as set forth in claim 5, wherein the second mentioned step is accomplished manually such that the degree of downward shift of said bands may be controlled by ear tuning of said downward shifts to obtain speech of maximum intelligibility.

7. Apparatus for improving the intelligibility of underwater speech communication for a deeply submerged diver whose voice is originally unintelligible because of frequency distortion due to ambient conditions, on the speaker, of lighter than air gaseous atmosphere and greater than atmospheric pressure comprising:

transducer means for converting said audio speech to an analog electrical signal,

lter means connected to the output of said transducer means for dividing said analog signal to selectively obtain at least two spaced bands of frequency within the range of 850 to 6,500 cycles having extremely sharp skirts containing a substantial portion of the original unintelligible speech,

frequency shifting means electrically connected to said lter means for linearly shifting said bands downwardly in frequency by respectively different degrees of shift, the higher band being shifted by a greater amount, dependent upon said ambient conditions and in relative degree such as to move, said bands closer together, but still spaced apart,

adder means connected to combine the output of said frequency shifted bands, and

second transducer means for converting the electrical output of said adder, means into an intelligible audio signal.

8. Apparatus as set forth in claim 7, said filter means and frequency shifting means being selected such that said bands after downward frequency shift remain suciently spaced in frequency to avoid intermodulation.

9. Apparatus as set forth in claim 7, said frequency shifting means including for each of said bands modulator and oscillator means for rst shifting said bands upwardly in frequency and demodulator and oscillator means for shifting said band frequencies downward, together with lter means for eliminating undesired frequencies.

10. Apparatus as set forth in claim 9, said oscillator means associated with said demodulator being manually adjustable such that the degree of downward shift of said bands may be controlled by ear tuning of said downward shifts to obtain speech of maximum intelligibility.

References Cited UNITED STATES PATENTS 4/1962 Schroeder 179-1555 12/1939 Riesz 179 OTHER REFERENCES Bell Laboratories Report, IEEE Spectrum, 5/ 67.

U.S. Cl. X.R. 179-15 .55

Images